AI-Generated Voice Mimicking Taiwan President Lai Ching-te Reportedly Used in Political Video Criticizing Government Food-Safety Response

A political video titled "It's Out of Your Control" reportedly used an AI-generated voice resembling Taiwan President Lai Ching-te while criticizing the government's response to a cooking-oil safety scandal. Police said the voice could be mistaken for Lai and later confirmed it was AI-generated. Opposition figures defended the video as political expression while prosecutors opened a forgery investigation.

Risk classification

• Primary risk domain: 4 Malicious actors
• Primary risk subdomain: 4.1 Disinformation, surveillance, and influence at scale

The incident involves the intentional creation and dissemination of an AI-generated deepfake of a country's leader to influence public opinion and political processes surrounding a food safety scandal.

Additional risk subdomains

• 3.1 False or misleading information: The AI cloned voice generated false statements attributed to the president, creating misleading content that could deceive the public.

Causal factors

• Entity: Human
• Intent: Intentional
• Timing: Post-deployment

The incident was caused by human political campaigners intentionally using a deployed AI voice cloning tool to generate a deepfake of the president's voice for political campaigning.

EU AI Act risk tier

• Risk tier: 3 Limited Risk

Risk Level 3: Limited Risk. The report describes the use of AI-generated deepfakes, which are subject to transparency and mandatory labeling obligations under the EU AI Act.

AI system and alleged parties

• AI system: unspecified
• AI purpose: Voice Generation; Deepfake Video Generation
• Behaviour type: Tool
• Alleged developer: Voice cloning technology developers, Synthetic media generation technology developers, Synthetic audio generation technology developers, Deepfake technology developers
• Alleged deployer: Wei Chun-yu, Synthetic media creators, Political operatives, Information manipulation actors in Taiwan, Information manipulation actors
• Alleged harmed parties: YouTube users, public figures, Political figures, Lai Ching-te, General public of Taiwan, General public, Epistemic integrity, Democratic integrity

Harm severity

Severity ratings are based on this scale.

Highest direct severity in any category: Minor. Severity is scored from Negligible to Catastrophic in each harm category, for harm the reports describe as caused directly or indirectly by the AI system.

• Physical: direct Negligible, indirect Negligible
• Infrastructure: direct Negligible, indirect Negligible
• Property: direct Negligible, indirect Negligible
• Financial: direct Negligible, indirect Negligible
• Environmental: direct Negligible, indirect Negligible
• Malicious content: direct Minor, indirect Minor
• Differential treatment: direct Negligible, indirect Negligible
• Civil rights: direct Negligible, indirect Minor
• Democracy: direct Negligible, indirect Minor
• Privacy: direct Minor, indirect Minor
• Psychological: direct Negligible, indirect Negligible
• Epistemic: direct Minor, indirect Minor
• Child sexual exploitation and abuse: direct Negligible, indirect Negligible

Malicious content

Reported: The report explicitly describes the creation of an AI-generated deepfake video mimicking the president's voice to criticize the government.

Directly caused: An AI-cloned voice of President William Lai was used in a video titled 'It's Out of Your Control' (油不得你) criticizing the government's handling of a cooking oil safety scandal.

Indirectly caused: Three additional AI-generated audio clips of Taipei Mayor Chiang Wan-an's voice were played at a news conference by DPP politicians as examples.

Civil rights

Reported: The report explicitly describes concerns over freedom of expression, state intimidation, and potential violations of personal data protection and personal rights.

Directly caused: The creator of the video, Wei, was questioned at his home by police, which KMT politicians characterized as state intimidation and an infringement on freedom of expression.

Indirectly caused: The investigation created a 'chilling effect' and raised concerns about selective law enforcement and the creation of a 'police state'.

Democracy

Reported: The report explicitly discusses how deepfake impersonations of government officials could undermine democratic decision-making and destabilize democratic elections.

Directly caused: The deepfake video of the president was released a day before a political rally, potentially misleading the public and disrupting public order.

Indirectly caused: The incident sparked intense political conflict between the DPP and KMT, with accusations of political persecution and selective law enforcement, eroding trust in public institutions.

Privacy

Reported: The report explicitly mentions potential violations of the Personal Data Protection Act and unlawful collection/processing of personal data (voice rights).

Directly caused: The AI cloned the voice of President William Lai without his consent, leading to legal complaints under the Personal Data Protection Act.

Indirectly caused: DPP politicians also generated and played voice clones of Taipei Mayor Chiang Wan-an, potentially violating his voice/privacy rights.

Epistemic

Reported: The report explicitly describes the potential for deepfake content to mislead the public into believing a fabricated statement was genuine.

Directly caused: The video simulated President William Lai's voice criticizing the government, which authorities warned could mislead the public into believing it was an official statement.

Indirectly caused: The spread of such deepfakes erodes public trust in facts and distorts public perception of official statements.

People affected

• Occurrences reported: 2
• People reportedly harmed: 2
• People reportedly exposed: 10000

Potential causes

Management

• Commissioning of AI Content: Political managers commissioned AI voice clones without safety oversight.
• Lack of Ethical AI Guidelines: Campaign teams lacked clear guidelines on the ethical use of AI.

Technology

• Realistic AI Voice Cloning: AI cloned the president's voice, making it sound highly authentic.
• High-Fidelity Deepfake Tools: Advanced AI tools allow easy generation of realistic fake audio.

Data Inputs

• Use of Public Audio for Training: Public voice recordings were used without consent to train the AI model.

Human Factors

• Intent to Mislead Viewers: Creators used AI to impersonate a leader to criticize the government.
• Low Public AI Literacy: General public struggles to distinguish cloned voices from real statements.

Process and Methods

• Inconspicuous AI Disclaimers: Disclaimers were hidden in the description box rather than on the video.
• Inadequate Content Verification: No robust process existed to verify the authenticity of political videos.

Regulatory Environment

• Inadequate AI Basic Act: Current framework lacks clear legal liabilities and penalty mechanisms.
• Narrow Deepfake Laws: Deepfake regulations are mostly limited to election-related fraud.
• No Mandatory Labeling Rules: No legal mandate exists to prominently label AI-generated political media.

Information quality

• Classification confidence: High
• Reason for confidence: The reports provide consistent and detailed coverage of the incident from multiple news sources, outlining the creation of the video, the police response, the political reactions, and the specific legal complaints filed by both sides.
• Ambiguities identified: It is slightly ambiguous whether the AI disclaimer was actively added by the creator or automatically by YouTube, and the exact AI tool used to clone the voices is not named.
• Alternative interpretations: The incident could be interpreted primarily as a political free speech dispute rather than an AI safety failure, though the central role of deepfake technology makes it an AI safety issue.

A political campaign in Taiwan used AI voice cloning to impersonate President William Lai criticizing the government, sparking police investigations and political disputes. While presenting minor national security risks, the incident highlights the evolving threat of deepfakes to democratic norms, domestic political stability, and civil liberties.

• Overall national security impact: Minor
• Scope: Single nation
• Primary target: Taiwan
• Alleged perpetrator: Wei Chun-yu

Threat characteristics

• Imminence: Near-term. The incident sparked immediate legal and political repercussions requiring monitoring and regulatory debates over the subsequent weeks.
• Autonomy: Human-controlled. The AI tool was directly utilized by human actors as a media-generation utility, with no autonomous agency or decision-making.
• Novelty: Evolved capability. Voice cloning is an established threat, but targeting a sitting president for domestic political messaging represents an advanced application of the technology.

Impact by dimension

• Physical security: Negligible. No physical security threats, kinetic targeting, or critical infrastructure impacts were reported in this incident.
• Information security: Minor. Involves AI-cloned voice of a national leader used in domestic political campaigns, mimicking official statements, though without evidence of foreign intelligence involvement.
• Sovereignty: Minor. Limited impact on core state authority, though it triggered investigations by the Criminal Investigation Bureau and heated debate over election integrity and police overreach.
• Economic security: Negligible. No significant threats to strategic industries, financial markets, or technological competitive advantages were identified.
• Societal stability: Minor. Prompted concerns over freedom of expression and a potential chilling effect due to police actions, but did not lead to widespread civil unrest.